The antigen binding sites of conventional antibodies are formed primarily by the hypervariable loops from both the heavy and the light chain variable domains. Functional antigen binding sites can however also be formed by heavy chain variable domains (VH) alone. In vivo, such binding sites have evolved in camels and camelids as part of heavy chain antibodies, which consist only of two heavy chains and lack light chains. Furthermore, analysis of the differences in amino acid sequence between the VHs of these camel heavy chain-only antibodies (also referred to as VHH) and VH domains from conventional human antibodies helped to design altered human VH domains (Lutz Riechmann and Serge Muyldermans, J. of Immunological Methods, Vol. 231, Issues 1 to 2, 1999, 25-38).
Similarly, it has been shown that by mutation studies of the interface residues as well as of the CDR3 on the VH of the anti-Her2 antibody 4D5 in parallel with the anti-hCG VHH H14, some mutations were found to promote autonomous VH domain behaviour (i.e. beneficial solubility and reversible refolding) (Barthelemy P A et al., 2008, J. of Biol. Chemistry, Vol 283, No 6, pp 3639-3654). It was also found that increasing the hydrophilicity of the former light chain interface by replacing exposed hydrophobic residues by more hydrophilic residues improves the autonomous VH domain behaviour. These engineered VHs were shown to be predominantly monomeric at high concentration, however low quantities of dimers and other aggregates of said engineered VHs were also found that presumably farm relative weak interaction similar to those described in the art for VL-VH pair interactions. Similarly, a camelized VH, called cVH-E2, is claimed to form dimers in solution in a concentration dependent manner i.e. at concentrations above 7 mg/ml (but note that data has not been shown in study; Dottorini et al., Biochemistry, 2004, 43, 622-628). Below this concentration, the dimer likely dissociates into monomers and it remains unclear whether these dimers were active (i.e. binding antigen).
Furthermore, it has recently been reported that a truncated llama derived VHH (the first seven amino acids are cleaved off) with a very short CDR3 (only 6 residues) called VHH-R9 forms a domain swapped dimer in the crystal structure. Since VHH-R9 has been shown to be functional in solution (low Kd against hapten) and to consist of a monomer only, it is likely that dimerization occurred during the very slow crystallization process (4 to 5 weeks) and that elements such as N-terminal cleavage, high concentration conditions and short CDR3 could lead or contribute to the “condensation” phenomena (see in particular also conclusion part of Spinelli et al., FEBS Letter 564, 2004, 35-40). Sepulveda et al. (J. Mol. Biol. (2003) 333, 355-365) has found that spontaneous formation of VH dimers (MD) is in many cases permissive, producing molecules with antigen binding specificity. However, based on the reported spontaneous formation (versus the dimers fanned by PIA reported herein) and the lack of stability data on the non-fused dimers, it is likely that these are weakly interacting dimers similar to the ones described by Barthelemy (supra).
Taken together, the literature describes the formation of dimers of single variable domains and fragments thereof that a) are interacting primarily on relatively weak hydrophobic interaction (which are e.g. depending on the concentration, reversible), and/or b) occur in another occasion only in the crystallisation process (e.g. as a result of crystal packing forces). Moreover, it has been described that these dimers were not binding antigens anymore (as in Spinelli (supra)) or it is unclear whether these dimers were binding dimers (as in Dottorini (supra) and Barthelemy (supra)).
It has been found (see e.g. WO 09/109,635) that stable dimer-complexes can be formed in solution with polypeptides comprising at least one single variable VHH domain. These dimer-complexes are also herein referred to as non-fused-dimers.